A Fuzzy-Based Dynamic Load-Balancing Algorithm

Many dynamic load-balancing algorithms have been proposed for parallel and discrete simulations. But the actual performances of these algorithms have been far from ideal, especially in the heterogeneous environment. In this paper, we design and implement a load-balancing system based on fuzzy logic control. The fuzzy algorithm has been implemented in a loosely coupled distributed system. On-line of workload measure has been addressed herein as being the load information policy, negotiation policy, and migration policy. The experimental results indicate that the fuzzy-based loadbalancing algorithm not only effectively reduces the amount of communication messages but also provides considerable improvement in overall performance such as short response times, high throughputs, and short turnaround times.

[1]  Tzung-Pei Hong,et al.  A load balancing algorithm using prediction , 1997, Proceedings of IEEE International Symposium on Parallel Algorithms Architecture Synthesis.

[2]  Walter H. Kohler,et al.  Models for Dynamic Load Balancing in a Heterogeneous Multiple Processor System , 1979, IEEE Transactions on Computers.

[3]  Cauligi S. Raghavendra,et al.  A Dynamic Load-Balancing Policy With a Central Job Dispatcher (LBC) , 1992, IEEE Trans. Software Eng..

[4]  Andrew A. Goldenberg,et al.  Development of a systematic methodology of fuzzy logic modeling , 1998, IEEE Trans. Fuzzy Syst..

[5]  Domenico Ferrari,et al.  An Empirical Investigation of Load Indices for Load Balancing Applications , 1987, Performance.

[6]  Lois M. L. Delcambre,et al.  Load balancing with network cooperation , 1991, [1991] Proceedings. 11th International Conference on Distributed Computing Systems.

[7]  Peter M. Dew,et al.  A periodic symmetrically-initiated load balancing algorithm for distributed systems , 1994, OPSR.

[8]  Edward D. Lazowska,et al.  A Comparison of Receiver-Initiated and Sender-Initiated Adaptive Load Sharing , 1986, Perform. Evaluation.